metadata
library_name: transformers
language:
- ro
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
datasets:
- IoanaLiviaPopescu/RealVoiceSynthVoice-400-2-Wavnet-B
metrics:
- wer
model-index:
- name: >-
IoanaLiviaPopescu/IoanaLiviaPopescu/real-data-synth-data-400-2-Wavnet-B-Wavnet-B-whisper-small-v0
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: IoanaLiviaPopescu/RealVoiceSynthVoice-400-2-Wavnet-B
type: IoanaLiviaPopescu/RealVoiceSynthVoice-400-2-Wavnet-B
config: default
split: test
args: 'split: validation'
metrics:
- name: Wer
type: wer
value: 17.88677853586576
IoanaLiviaPopescu/IoanaLiviaPopescu/real-data-synth-data-400-2-Wavnet-B-Wavnet-B-whisper-small-v0
This model is a fine-tuned version of openai/whisper-small on the IoanaLiviaPopescu/RealVoiceSynthVoice-400-2-Wavnet-B dataset. It achieves the following results on the evaluation set:
- Loss: 0.3999
- Wer: 17.8868
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 32
- eval_batch_size: 16
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 0 | 0 | 0.6024 | 27.8812 |
| 0.3857 | 1.0 | 26 | 0.4413 | 19.5095 |
| 0.1796 | 2.0 | 52 | 0.3877 | 17.9974 |
| 0.1036 | 3.0 | 78 | 0.3856 | 18.1818 |
| 0.0699 | 4.0 | 104 | 0.3975 | 17.9237 |
| 0.0539 | 5.0 | 130 | 0.3999 | 17.8868 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1